Table 3 Anomaly detection accuracy with the PRONOSTIA dataset (%)
From: A Dual-Task Learning Approach for Bearing Anomaly Detection and State Evaluation of Safe Region
Group number | Bearing number | RMS | NESES | Auto-encoder | Proposed method |
|---|---|---|---|---|---|
1 | 1_1 | 96.4 | 91.2 | 96.2 | 98.2 |
1_2 | 95.8 | 50.0 | 95.8 | 100.0 | |
1_3 | 98.3 | 84.2 | 98.3 | 99.3 | |
1_4 | 99.1 | 62.6 | 98.0 | 99.2 | |
1_5 | 92.3 | 67.3 | 100.0 | 98.9 | |
1_6 | 90.6 | 50.0 | 100.0 | 100.0 | |
1_7 | 100.0 | 78.0 | 100.0 | 99.5 | |
2 | 2_1 | 94.7 | 50.0 | 97.4 | 98.9. |
2_2 | 100.0 | 64.3 | 100.0 | 100.0 | |
2_3 | 100.0 | 85.0 | 99.2 | 99.0 | |
2_4 | 100.0 | 100.0 | 100.0 | 100.0 | |
2_5 | 100.0 | 85.0 | 98.3 | 99.4 | |
2_6 | 100.0 | 100.0 | 100.0 | 100.0 | |
2_7 | 100.0 | 87.5 | 100.0 | 100.0 | |
3 | 3_1 | 80.8 | 50.0 | 96.2 | 98.8 |
3_2 | 97.5 | 50.0 | 100.0 | 98.5 | |
3_3 | 85.7 | 54.8 | 100.0 | 100.0 | |
Average | 0.96 | 0.69 | 0.98 | 0.99 | |